Re: [Axiom-developer] Aldor extend for extending Axiom code

[Ralf Hemmecke]

> The second restriction is more severe. In fact, I think it's a show
> stopper. The current signature of Plus is
>
> Plus(
>     F: (T: LabelType) -> CombinatorialSpecies(T),
>     G: (T: LabelType) -> CombinatorialSpecies(T)
> )(L: LabelType): CombinatorialSpecies(L) == add { ... }
>
> Although this compiles and loads well, we get 
>
> )sh Plus
>  Plus(F: (LabelType -> CombinatorialSpecies T),G: (LabelType ->
>  CombinatorialSpecies T))  is a domain constructor
>  Abbreviation for Plus is PLUS 
>  This constructor is exposed in this frame.
>  Issue )edit species.as to see algebra source code for PLUS 
>
> ------------------------------- Operations --------------------------------
>  Plus(F,G) is a domain constructor.
>  Abbreviation for Plus is PLUS 
>  This constructor is exposed in this frame.
>  Issue )edit species.as to see algebra source code for PLUS 
>
> ------------------------------- Operations --------------------------------
>
>
> That's it. In fact, we cannot even really try to use it:
>
> (5) -> Plus(SingletonSpecies, SingletonSpecies) IntegerLabel
>  
>    Although SingletonSpecies is the name of a constructor, a full type 
>       must be specified in the context you have used it. Issue )show 
>       SingletonSpecies for more information.
>
> depressing.

Well, you are pretty far. Congratulations.

I have a feeling why

Plus(SingletonSpecies, SingletonSpecies) IntegerLabel

does not work. It is probably for the same reason that you cannot put 
elements of type (T: LabelType) -> CombinatorialSpecies(T) into a list.

Maybe a rewrite with your LabelSpecies and "coerce" would do the job.

But that is ugly!!!

*We desparately need an open aldor compiler that can help the Axiom 
interpreter.*

Ralf